Kalman Filter (pykalman) with two measurements for the same variable

Question

in the last days I was using pykalman to correct GPS coordinates with acceleration and velocity measurements, which is working fine. Now I would like to combine the resulting coordinates (1) with a second coordinate measurement (2) which is very precise but has a much lower logging rate. I masked all elements of the second measurement which have no value (== 0).

My problem is, that I don't know how the matrices have to look like. Maybe the Kalman Filter is not even suited for what I want to do.

Thank you very much in advance.

for i in range(len(lat1)):

    if i == 0:
        measurements = ma.asarray([(lat1[i], long1[i], lat2[i], long2[i])])
    else:
        measurements = ma.append(measurements, [[lat1[i], long1[i], lat2[i], long2[i]]], axis=0)


for i in range(len(measurements)):
    if measurements[i, 2] == 0:
        measurements[i, 2] = ma.masked
    if measurements[i, 3] == 0:
        measurements[i, 3] = ma.masked


initial_state_mean = ma.asarray([measurements[0, 0], measurements[0, 1], measurements[0, 2], measurements[0, 3]])

transition_matrix = [[0, 0, 1, 0],  
                     [0, 0, 0, 1],  
                     [0, 0, 0, 0],  
                     [0, 0, 0, 0]]  


observation_matrix = [[1, 0, 0, 0],
                      [0, 1, 0, 0],
                      [0, 0, 1, 0],
                      [0, 0, 0, 1]]


kf1 = KalmanFilter(transition_matrices=transition_matrix,
                   observation_matrices=observation_matrix,
                   initial_state_mean=initial_state_mean,
                   observation_covariance=[[0.1, 0, 0, 0],
                                           [0, 0.1, 0, 0],
                                           [0, 0, 1, 0],
                                           [0, 0, 0, 1]],

                   transition_covariance=[[0, 0, 1, 0],
                                          [0, 0, 0, 1],
                                          [0, 0, 0, 0],
                                          [0, 0, 0, 0]]
                   )


(smoothed_state_means, smoothed_state_covariances) = kf1.smooth(measurements)

lat_corr = smoothed_state_means[:, 0]
long_corr = smoothed_state_means[:, 1]